extracting-sdoh

Extract social determinants of health and map findings to ICD-10-CM Z-codes.

5.0k|615|Updated Oct 4, 2025
One-click install
npx skills add https://github.com/maziyarpanahi/openmed --skill extracting-sdoh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: extracting-sdoh
Source: https://github.com/maziyarpanahi/openmed/tree/main/skills/extracting-sdoh
Command: npx skills add https://github.com/maziyarpanahi/openmed --skill extracting-sdoh

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill recovers social determinants of health documented in clinical narratives but often missing from structured coding, then suggests relevant ICD-10-CM Z-codes for human confirmation.

Core Features & Use Cases

  • SDOH Extraction: Identify housing instability, food insecurity, unemployment, transportation barriers, social isolation, financial strain, and related social risks in clinical text.
  • Z-Code Mapping: Map extracted spans to candidate ICD-10-CM Z55–Z65 codes using a curated lookup aligned with Gravity Project domains.
  • Safety and Interoperability: Preserve source spans, confidence scores, negation and temporality considerations, and support downstream FHIR Condition or Observation workflows without auto-billing inferred codes.
  • Use Case: Use the Skill to surface undocumented housing and transportation risks from de-identified patient notes for health-equity dashboards, coder review, or closed-loop referrals.

Quick Start

Use the extracting-sdoh skill after de-identifying a clinical note and running OpenMed NER to produce span-anchored SDOH findings and candidate ICD-10-CM Z-codes for human review.

Frequently Asked Questions about extracting-sdoh

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract social determinants of health from clinical narratives?

To extract social determinants of health from clinical narratives, this Skill identifies housing instability, food insecurity, and transportation barriers from text and maps them to candidate ICD-10-CM Z-codes. It requires de-identified clinical notes and OpenMed NER output to function.

Can I map clinical text findings to ICD-10-CM Z-codes for health equity analytics?

Yes, you can map clinical text findings to ICD-10-CM Z-codes for health equity analytics. The Skill uses a curated lookup aligned with Gravity Project domains to map extracted spans to candidate Z55–Z65 codes for human review.

Does extracting SDOH require de-identified clinical notes and NER output?

Yes, extracting SDOH requires de-identified clinical notes and OpenMed NER output. The Skill relies on span offsets, confidence scores, and NER findings to accurately surface social risks and suggest candidate Z-codes for human validation.

How do I generate FHIR Conditions from extracted SDOH Z-codes?

To generate FHIR Conditions from extracted SDOH Z-codes, the Skill preserves source spans and confidence scores to support downstream FHIR Condition or Observation workflows. It provides candidate codes for human validation without auto-billing inferred Z-codes.

What are the limitations of using automated SDOH extraction for medical coding?

A key limitation of automated SDOH extraction for medical coding is that it cannot auto-bill inferred codes. The Skill requires human validation for negation, temporality, subject context, and billing safety before applying suggested ICD-10-CM Z-codes to patient records.

When do I need Z-codes for risk stratification and closed-loop referrals?

You need Z-codes for risk stratification and closed-loop referrals when clinical narratives document undocumented social risks. This Skill surfaces housing and transportation risks from de-identified notes, providing candidate Z-codes for coder review and referral workflows.